Deep learning technology has brought convenience and advanced developments but has become untrustworthy due to its sensitivity to adversarial attacks. Attackers may utilize this sensitivity to manipulate predictions. To defend against such attacks, existing anti-adversarial methods typically counteract adversarial perturbations post-attack, while we have devised a proactive strategy that preempts by safeguarding media upfront, effectively neutralizing potential adversarial effects before the third-party attacks occur. This strategy, dubbed Fast Preemption, provides an efficient transferable preemptive defense by using different models for labeling inputs and learning crucial features. A forward-backward cascade learning algorithm is used to compute protective perturbations, starting with forward propagation optimization to achieve rapid convergence, followed by iterative backward propagation learning to alleviate overfitting. This strategy offers state-of-the-art transferability and protection across various systems. With the running of only three steps, our Fast Preemption framework outperforms benchmark training-time, test-time, and preemptive adversarial defenses. We have also devised the first, to our knowledge, effective white-box adaptive reversion attack and demonstrate that the protection added by our defense strategy is irreversible unless the backbone model, algorithm, and settings are fully compromised. This work provides a new direction to developing proactive defenses against adversarial attacks.
翻译:暂无翻译